BACHELOR'S THESIS
Fast cook-off test with a sand bed burner
Evaluation of the heating process with LPG compared to Jet A1
Björn Evers Peter Möllerström
2013
Fire Protection Engineering Fire Protection Engineer
Luleå University of Technology
Department of Civil, Environmental and Natural Resources Engineering
Fast cook-off test with a sand bed burner
- Evaluation of the heating process with LPG compared to Jet A1
Luleå University of Technology The program of Fire Protection Engineering
Department of Civil, Environmental and Natural resources engineering Supervisors: Ulf Wickström, LTU and Alf Prytz, Saab Bofors Test Center AB
iii
Preface
This is a thesis from the fire protection engineer program on Luleå University of Technology. The work has been performed with support from Bofors Test Center AB, SP Technical research institute of Sweden (SP) and Luleå University of Technology (LTU).
Special thanks are directed to Alf Prytz, Jon Toreheim and Kjell Sånebo with crew at Bofors Test Center for the help and support during the test days and answers on all our questions during the project.
We also thank Alexandra Byström for input on the construction of the probe and planning for the test days, Lars Åström and Mats Petersson at the workshop Complab at LTU for help with the construction and validation of the probe, Michael Magnusson at SP for giving us a tour at the testing facility in Borås and our supervisor Ulf Wickström for help and support during the whole project.
Luleå, Sweden, June 2013
Björn Evers Peter Möllerström
v
Abstract
Many modern munitions are designed to be insensitive to outside influences. To classify a munition as insensitive certain tests must be accomplished. One of these tests is a fast cook-off test to simulate the munitions behavior in an intense fire. According to the NATO standard, STANAG 4240, the test must be performed with jet fuel, for example Jet A1. Due to high testing costs and the environmental impact of the standard test, attempts have been made to replace jet fuel with alternative fuels. The aim of this report is to construct a sand bed burner based on propane and compare the heating process with a Jet A1 pool fire. The report also presents an alternative measurement method to ensure the same thermal impact as in a Jet A1 pool fire.
The sand bed burner is based on letting propane diffuse through a sand bed. This will make the combustion result in flames with a higher radiation than premixed flames. A probe measuring steel surface temperature and gas temperature was constructed to make the comparison possible. The steel temperature was then approximated as the adiabatic surface temperature.
Three different tests were made; with the sand bed burner, with a Jet A1 pool fire and with the existing Bofors Test Center (BTC) gas burner. The results were then analyzed by comparing the thermal impact in each direction.
The result shows some differences in the heating pattern. Jet A1 provides the most even heating but the maximum gas temperatures are lower than with the other methods. The BTC gas burner shows the most uneven heating, with over 400 ˚C difference from the lower to the upper point. The sand bed burner is somewhere between the other two methods.
The two main problems were identified during the tests with the sand bed burner. No combustion occurred in the middle of the burner and the gas supply varied due to cooling of the tubes when the LPG evaporated.
Based on the results from the test carried out it can be assumed that with a slightly modified sand burner and an adequate gas supply further test in larger test are likely to be successful in terms of yielding similar thermal exposure as the test specified in the standard STANAG 4240.
Keywords: Insensitive munitions, IM, IM-testing, fast cook-off test, FCO, adiabatic surface temperature, TAST, sand bed burner, propane, LPG, Jet A1, STANAG 4240, AOP-39
vii
Nomenclature
Symbols m = Mass [kg]
C = Heat capacity [J/K]
c = Specific heat capacity (C/m) [J/kg K]
U = Internal energy [J]
u = Internal energy per mass unit [J/kg]
h = Heat transfer coefficient [W/m2K]
ε = Emissivity
K = Effective emissivity coefficient [m-1] q = Heat [Ws = J]
= Combustion efficiency = Heat of combustion [J/kg]
= Scaling factor for calculating weight loss
σ = Stefan Boltzmann constant, [W/m2K4] T = Temperature [˚C or K]
τ = Time constant [s]
α = absorptivity Superscripts
” = per unit area
˙ = per unit time Subscripts emi = emitted fl = flame = ambient
AST = Adiabatic Surface Temperature s = surface
r = radiative g = gas c = convective tot = total inc = incident abs = absorbed rad = net radiation con = convection TC = thermocouple
viii Definitions
Adiabatic surface temperature – The equilibrium temperature of a surface which does not absorb or lose any energy, i.e. ̇ 1.
STANAG – North Atlantic Treaty Organization Standardization Agreement2. AOP – North Atlantic Treaty Organization Allied Ordnance Publication3.
Insensitive Munitions (IM) – “Munitions which reliably fulfill their performance, readiness and operational requirements on demand and which minimize the probability of inadvertent initiation and severity of subsequent collateral damage to weapon platforms, logistic systems and personnel when subjected to selected accidental and combat threats”4.
LPG – Liquefied petroleum gas, such as propane.
1 Wickström, Heat transfer in fire technology, 49
2 STANAG 4240
3 AOP-39
4 AOP-39, Annex A
ix
Table of contents
1 INTRODUCTION... 1
1.1 Background ... 1
1.1.1 Bofors Test Center ... 1
1.1.2 Insensitive munitions ... 1
1.1.3 Fast cook-off test ... 1
1.1.4 Earlier studies ... 2
1.1.5 The advantages of LPG against Jet A1 fuel ... 3
1.2 Purpose and objective ... 4
1.2.1 Purpose ... 4
1.2.2 Objective ... 4
1.3 Limitations... 4
2 METHOD ... 5
2.1 Initial method ... 5
2.2 Test method ... 5
2.2.1 Test of the TAST-probe... 5
2.2.2 Small scale Jet A1-test ... 6
2.2.3 Sand bed burner test ... 6
2.2.4 Test with the existing LPG system ... 6
3 THEORY... 8
3.1 Theory of heat transfer ... 8
3.1.1 General theory of heat transfer... 8
3.1.2 Heat transfer for the sand bed burner ... 10
3.2 Measuring Tr, Tg and TAST ... 15
3.2.1 Estimating the heat transfer coefficient and the surface emissivity ... 16
3.3 Difference between Jet A1 and propane flame ... 18
4 RESULTS ... 20
4.1 Sand bed burner ... 20
4.1.1 Construction of the sand bed burner... 20
4.2 Adiabatic surface temperature probe ... 21
4.2.1 Construction of the probe ... 21
4.2.2 Test of the probe ... 23
4.3 Test results ... 25
x
4.3.1 Sand bed burner test ... 25
4.3.2 Small scale Jet A1-test ... 29
4.3.3 Test with the existing LPG system ... 30
5 ANALYSIS ... 32
5.1 Analysis of the test of the probe ... 32
5.2 Analysis of the test results ... 33
5.2.1 Comparing the 3 different heating processes ... 33
5.2.2 Calculation of the incident heat flux and radiation temperature... 37
5.2.3 Analyzing variations in the gas flow ... 40
6 DISCUSSION... 42
7 CONCLUSIONS ... 48
7.1 Suggested test method ... 48
8 REFERENCES ... 51 APPENDIX A. DESIGN DRAWINGS FOR THE SAND BED BURNER
APPENDIX B. DESIGN DRAWINGS OF THE TAST-PROBE APPENDIX C. MISSION STATEMENT IN SWEDISH
APPENDIX D. PROPANE FLOW FOR SAME HRR AS FOR JET A1 APPENDIX E. PRODUCT SPECIFICATIONS FOR THERMOCOUPLES
1
1 Introduction 1.1 Background
1.1.1 Bofors Test Center
In year 1886 the first firing range in Karlskoga was built by Bofors5. Bofors Test Center (BTC) is now a facility for testing of both military and civilian products with core business of testing products with explosive contents6. The facility is owned by Saab Bofors Test Center AB, which is owned by Saab Dynamics, BAE Systems and Eurenco.
The test area is 200 km2 and has several different firing ranges and test facilities. On the facilities, realistic environments can be created that may expose products to different types of impacts. The reaction of the products are analyzed and documented in order to draw conclusions about whether the product meets the requirements placed on it or not.
1.1.2 Insensitive munitions
In 1967 a plane accidentally fired a missile on the flight deck of USS Forrestal aircraft carrier. This lead to a fire which spread to other planes and lead to a chain reaction of explosions. This was the starting point of developing munitions with less sensitivity to outside influences, such as fires. The insensitive munitions (IM) are developed to withstand various kinds of external influences. The IM- tests are based on six parts7, each regulated in international standards; bullet impact, fragment impact, shaped charge jet impact, fast cook-off, slow cook-off and sympathetic reaction. Each test evaluates the munitions resistance to the threat. This report focuses on an alternative way to design the fast cook-off test.
1.1.3 Fast cook-off test
The aim of the fast cook-off test (FCO-test) is to assess the response of munitions and weapon systems exposed to a rapid thermal impact8. The design of the weapon should maximize the time to reaction and minimize the intensity of the reaction. The goal of the test is to determine how the weapon system reacts over time when exposed to heat. The test is designed to simulate the worst case scenario with the fastest heating possible in fire situations. The goal for the munition is to have no worse reaction than burning9.
During the test the entire object will be surrounded in flames and the objects reaction over time will be analyzed. The flames should be fuel-rich and the heat transfer to the object should be about 90%
radiative. The average temperature of the flames should be at least 800 ˚C and the flame temperature should reach 550 ˚C in less than 30 seconds.
5 Bofors Test Center. Historik.
6 Bofors Test Center. Nutid.
7 Bofors Test Center. Insensitive Munitions (IM)
8 STANAG 4240, 1-3.
9 AOP-39, 2.
2 The standard test is performed with jet fuel (Jet A1 or similar) but due to high costs, pollution and work environment an alternative test with LPG is considered. The AOP-39 standard opens up for alternative fuels if the environmental regulations ban use of hydrocarbon fire test10.
The requirements for the test according to STANAG 4240 and AOP-39 are summarized as:
- The flame temperature must reach 550 ˚C in 30 seconds and have an average temperature of at least 800 ˚C
- The area of the burner must be at least 1 m on each side of the object - The radiation must be dominant (approx. 90%)
- The object must be fully enclosed in flames
- The temperatures are measured with at least 4 thermocouples (TC) with at least one measurement every 5th second (0.2 Hz)
- The tests must not be performed in rain, snow or similar and the wind speed must not exceed 2.8 m/s
1.1.4 Earlier studies
Studies in replacing Jet A1 with LPG in the fast cook off tests are relatively new. A few test facilities are built. In United States a facility with LPG was built as a big cylinder with a fan11. The gas is injected in the airflow and a jet flame is created. The test showed that the difference in time to reaction was about 15% when comparing liquid fuel tests and LPG tests. The tests were also shown to be repeatable with less than 4 seconds in difference for time to reaction.
Figure 1. The existing BTC gas burner.
Bofors Test Center has a LPG system consisting of 14 standard propane burners each providing a premixed flame12, see Figure 1. Tests show that the temperatures in STANAG 4240 and AOP-39 are
10 AOP-39, Appendix 7 to Annex H.
11 Ford, Sub-scale Fast Cook-off Test Results
12 Toreheim, Fast Cook-Off Using Liquefied Propane Gas
3 reached with the measurement method prescribed in the standard (STANAG 4240). The time to reaction is also comparable with liquid pool fire tests. Problems with the test is that the radiative heat flux is low (which is not acceptable according to STANAG 4240) and the single flames create hot spots on the test item.
In summary, it seems unclear what is actually measured during the various tests. This makes it hard to evaluate and compare different systems. Many of the LPG projects are in the development phase.
There seems to be a need for a standardized measurement method, even during the standard liquid pool fire tests. As the standard looks today, it is only the number of thermocouples, location of them and the measurement time which is regulated. No thermocouple diameter is specified in the standard, something which is important to know when evaluating the temperature measured during the test (see chapter 3.2).
1.1.5 The advantages of LPG against Jet A1 fuel
There are many advantages to use LPG instead of Jet A1. Not only economic, but also health, environment and time consumption for the tests.
The LPG burns more efficiently due to its chemical composition, the chemical formula for LPG of propane is C3H8 and for kerosene (which is the main content in Jet A1) it is CnH2n+2 where n varies between 11 and 14. This gives that kerosene and Jet A1 needs more oxygen for an ideal (efficient) combustion. The LPG burn much cleaner than kerosene and gives cleaner residues.
Below some different properties are listed.
Environment:
Propane: No known environmental impacts13.
Jet A1: It is a slow degradable substance that partially dissolves in water, the substance can penetrate the soil and reach groundwater14. The hydrocarbons can be adsorbed onto organic materials in the soil and sediment. Larger spills can cause acute death of fish and other aquatic organisms, but is also harmful to birds and vegetation15.
Price:
Even though the price per kilogram may be higher for propane there are some economic advantages.
The possibility to abort the test will make the fuel consumption as low as possible. The short set up times for the tests will make it more cost efficient to perform the tests, more tests can be done in one day. There is also a possibility to in an easy way perform tests in smaller scale.
Work environment:
Propane: There are no known toxicological effects16. Propane is denser than air and can thereby be stifling.
Kerosene fractions: Mildly irritating to the skin and eyes, repeated or prolonged contact irritates the skin17. Prolonged exposure of kerosene vapor has produced species-specific renal damage on male
13 AGA, Säkerhetsdatablad Gasol
14 St1. Säkerhetsdatablad Jet A1
15 Fisher Scientific. Kerosene
16 AGA. Säkerhetsdatablad Gasol
4 rats. Experiences from humans have shown that vapors can irritate the eyes and respiratory tract.
Prolonged or repeated contact irritates and dries the skin. If swallowed it irritates the digestive tract.
Aspiration can cause fatal chemical pneumonia.
The ability to abort the test:
At large scale pool fire tests it is hard to extinguish a fire with water, and with foam the fuel becomes unusable. The remaining fuel can be difficult to collect and re-use with evaporation as a result. To let the remaining fuel burn or evaporate is both a big waste and an unnecessary cost. With a sand bed burner and propane, a test is easily stopped since the gas flow is controlled with shut-off valves.
1.2 Purpose and objective
1.2.1 Purpose
The purpose of this work is to develop and evaluate a small scale fast cook-off test with a LPG sand bed burner and compare it to a Jet A1 pool fire. Then determine the possibility of future use of a large scale burner in IM testing instead of the standard Jet A1 pool fire test.
1.2.2 Objective
The objective of the work is to construct a prototype of a sand bed burner and with a TAST-probe compare the LPG fire with the Jet A1 pool fire. A test with the existing BTC gas burner is made as a reference.
Questions to be answered:
- What similarities and differences are there in the heating process?
- What theoretical differences are there in thermal radiation and convection between an LPG fire and a Jet A1 pool fire?
- How shall the measurements with the probe be performed to validate the method during real tests?
1.3 Limitations
The limitations of this work are:
- It only looks at what the object is exposed to by the environment, not what is happening inside the object, such as thermal conductivity, expansion, etc.
- It only describes the similarities and differences between heating with propane and Jet A1, it will not optimize the model to minimize the difference.
- The tests are only made in small scale in an attempt to analyze whether it is possible to do tests on a large scale.
-
The tests are limited by the existing LPG system on Bofors Test Center. The gas flow will therefore be reduced to the maximum flow of the existing gas system.
17 St1. Säkerhetsdatablad Jet A1
5
2 Method
2.1 Initial method
This work is a part of a development process at BTC where in contact with Ulf Wickström the ideas of a sand bed burner as a replacement of the Jet A1 pool fire emerged. This led to a thesis where the principle of the plate thermometer where used to construct a round probe to measure the temperatures in a Jet A1 and a propane flame. Most of the theory is based on Wickströms DRAFT
‘Heat transfer in fire Technology’.
The project started with information gathering and literature studies to get an idea of the problems in the work ahead, and for the theory of the measurements. Also the standards18 being followed in the execution for the pool fire test was studied because they have to be taken in consideration in the development of the burner.
When all information needed were gathered, the development of the probe and the burner where started. Once started, it was clear that some more information was needed and more studies were performed. Discussions about the probe where done with Wickström, Alexandra Byström, SP Technical research institute of Sweden (SP), the workshop Complab at Luleå University of Technology (LTU) and with Pentronics. The probe was thereafter designed. The design of the sand bed burner where derived mostly in discussions with Wickström.
2.2 Test method
The purpose of the tests is to see the similarities and differences between the new propane burner and the standard test method (with Jet A1). At BTC, tests should be performed with the three different heating processes and then analyze the results of the measurements. Tests shall be performed with the current LPG burner (BTC gas burner), a smaller 1.0 x 1.15 m standard test fire (Jet A1-fire) and two tests with the sand bed burner.
The goal of the new LPG system is to show that the gas temperature reaches 550 ˚C in less than 30 seconds, the average gas temperature is 800 ˚C and to compare the difference in radiation between the sand bed burner and the Jet A1 pool fire.
Factors that need to be controlled during the test of the sand bed burner is gas flow, as it controls the heat release rate of the burner, wind speed, ambient temperature and the location of the probe.
Before the tests on site the TAST-probe must be assured to have the necessary quality. This control is made in the laboratory at LTU and the aim is to see that the thermocouples work properly when heated, and that the result is sufficiently reliable.
2.2.1 Test of the TAST-probe
To verify that the probe works with sufficiently high reliability a simple function test is performed.
The test shall be performed with controllable input to verify that the output given by the probe is sufficiently reliable.
The probe must be insulated and complete (see chapter 4.2.1) with welded 0.8 mm thermocouples for the steel surface temperature (approximated as the adiabatic surface temperature, AST) and 0.25
18 AOP-39 and STANAG 4240
6 mm thermocouple satellites for measurement of the gas temperature. A standard plate thermometer is used as a reference.
During the verification test the probe was heated according to the ISO 834-curve19 up to a maximum temperature of about 800 ˚C. The temperature was then maintained for at least 3 minutes. Output from the probe and the plate thermometer was compared with the input temperature.
2.2.2 Small scale Jet A1-test
The aim with the small scale Jet A1-test is to provide basic values of steel surface temperature and gas temperature in order to make a comparison with the new LPG system.
To simplify scaling a test tub with the same base area as the sand bed burner is used; 1.0 x 1.15 m2. Sufficient amount of fuel is added so the fire will burn for 5 minutes. Simplified Jet A1 can be expected to burn with 7 mm/min regardless of the size of the fire20. This gives that the total height of Jet A1 in the test must be 35 mm which equals 40.25 liters.
Measurements in the plume is done with the TAST-probe and with the coarse thermocouples previously used by BTC in their measurements. The probe provides the steel surface temperature and the gas temperature. The former coarse thermocouples are used as reference. The placement of the probe corresponds to the height used previously in similar tests. The values are recorded and compared to the other tests done with the sand bed burner.
2.2.3 Sand bed burner test
The aim of the test with the sand bed burner is to provide values of steel surface temperature and gas temperature in order to compare these with the values from the small scale Jet A1-test.
The needed gas flow during the tests was calculated to approx. 150 kg/s, see Appendix D. During this test the gas flow is limited by the existing LPG system on the test site. The system provides gas from 10 LPG bottles connected to a regulator which supplies the sand bed burner with gas through the inlet on the side of the burner.
The first step is to determine the most effective height of the probe. The temperatures in 4 different heights are measured with coarse thermocouples for 60 seconds. The probe should be placed where the highest temperature occurs, but it must not be placed above the flame. In the later tests, the probe must be fully enclosed in flames.
Measurements in the plume are done as similar as possible to the measurement in the small scale Jet A1-test. The probe is used for steel surface temperature and gas temperature and course 3 mm thermocouples as a reference. The values are then compared with the Jet A1-test.
During the tests the gas flow must be recorded in order to calculate the actual heat release rate. The tests are also recorded with camera.
2.2.4 Test with the existing LPG system
The BTC gas burner provides heat to the object with 14 premixed propane burners. The gas flow is assumed constant during the tests.
19 T = Ti + 345*log(8*t + 1) where t is time in minutes
20 STANAG 4240
7 A test with the existing LPG system is performed as a reference test and is compared both with the small scale Jet A1-test and with the sand bed burner test. The test is performed with the probe and the coarse thermocouples as similar to the tests above as possible. The values are then compared.
During the test the gas flow must be recorded to be able to calculate the actual heat release rate.
8
3 Theory
3.1 Theory of heat transfer
3.1.1 General theory of heat transfer
Heat transfer is based on the first law of thermodynamics; energy cannot be created or destroyed, only transformed21. Heat flux can be derived from the change in total energy of the system. The total energy is defined by the sum of the internal energy (U), potential energy (PE) and kinetic energy (KE):
E = U + PE + KE (3.1)
Heat transfer does not result in any change of the kinetic and potential energy. This gives that E = U.
If no work is done on the system, the first law of thermodynamics gives:
(3.2)
The specific heat capacity is the amount of energy required to heat a substance of a specific mass one degree Celsius and is defined by22:
⇒ (3.3)
Heat transfer to a definite volume is thereby driven by differences in temperature and is given by the integral23:
∫ where
For large temperature differences the specific heat capacity cannot be assumed constant over temperature24. Using the equation above, the heat supplied to an object can be calculated if the initial temperature, final temperature, density and the equation for the specific heat capacity are known. Furthermore, the heat transfer occurs in three ways25; radiation, convection and conduction.
Each part of the heat transfer theories will be presented in detail later.
3.1.1.1 Radiation
Radiation is heat transfer by electromagnetic waves26. The radiation emitted from a surface is given by the Stefan-Boltzmann law:
̇ (3.5)
21 Çengel, Thermodynamics: An Engineering Approach, 118-119.
22 Ibid, 128-133.
23 Wickström, Heat transfer in fire technology, chapter 3.1.
24 Atkins, Physical Chemistry , 55.
25 Wickström, Heat transfer in fire technology, chapter 3.1.
26 Ibid, 75.
c = Specific heat capacity (3.4) ρ = density
T1 = Initial temperature T2 = Final temperature
9 where ε is the surface emissivity and σ is the Stefan-Boltzmann constant (5.67*10-8 W/m2K4).
Emissivity is a description of how much energy that is emitted from a surface. For a perfect black body which does not emit any energy at all, the emissivity is 1. Other surfaces that will emit energy, the emissivity will vary between 0 and 1. With the same principles as above, the radiation from a flame is given by:
̇ (3.6)
In many cases, it is more interesting to know the amount of radiative energy absorbed by an object.
To know this, analyzes have to be made of how much of the radiation that will be reflected and emitted. For a grey body, α is defined as absorptivity, which is equal to the emissivity, ε. Therefore, we get α ≈ ε. The absorbed radiation is then calculated as:
̇ ̇ ̇ (3.7)
The remaining radiation will then be reflected away from the surface of the object. This gives ̇ ̇ . An expression for the total heat transfer by radiation to an object involved in a flame can now be summarized to:
̇ ̇ ̇ ⇒ ̇ ( ̇ ) ̇ (3.8) 3.1.1.2 Convection
Heat transfer by convection is heat that is transported by some form of fluid to a surface or an object27. Convection occurs due to thermal buoyancy generated by the difference in density that occurs when a fluid is heated. Convection can both have a warming effect (on a colder surface) or a cooling effect (on a hotter surface). Heat transfer between the fluid and a surface or an object can therefore be calculated with an apparently relatively simple formula:
̇ ( ) (3.9)
where hc is the convective heat transfer coefficient and Tg and Ts are the gas and surface temperatures respectively. In order to estimate the heat transfer coefficient calculations in several steps is required and this is the biggest engineering challenge due to heat transfer by convection.
Chapter 3.2.1.1 explains more about how an estimate of the convective heat transfer coefficient can be made.
3.1.1.3 Conduction
Conduction will only be presented briefly because of the heat transfer by conduction is assumed to be negligible in the tests. No estimates of conductivity will therefore be made in the report.
Conduction is heat transfer in a material due to increased molecular activity28. Temperature is a measure of the movement in and between atoms. When the activity of the material increases, so does the temperature. In this way, the heat is conducted in to the material.
Thermal conductivity calculations are different if the process has reached steady state or not. In the simplest case steady state can be assumed and the conduction is linear and in one dimension. In such
27 Wickström, Heat transfer in fire technology, 59.
28 Ibid, 14.
10 cases an analytical solution is possible. In many real cases, however, the heat conduction is non- linear and multidimensional. Most of the material parameters change with temperature and also the heat flux supplied often varies with time. In these cases, an analytical solution is virtually impossible and numerical methods must be applied.
3.1.2 Heat transfer for the sand bed burner
The overall heat transfer to the object is as given above, also see Figure 2:
[
] [
] [
] [ ] ̇ ̇ ̇ ̇ where ̇ ̇ ̇ ̇ (3.10) With the previously presented equations we get:
̇ ( ) ( ) ̇ ̇ (3.11)
Figure 2. Theoretical description of heat transfer to an object.
With this formula it is assumed that steady state has occurred and that the variation of the flame temperature, surface temperature and the gas temperature is not significant. In the experiments it’s assumed that the heat supplied to the object by conduction is negligible. To determine the amount of incoming radiation that is transferred to the object, the flame and surface emissivity (εfl and εs)
11 must be estimated. The emissivity of different surfaces is tabulated in various sources or it can be determined by experiments. For a small fire, flame emissivity can be calculated29:
where
This formula assumes that it is the same temperature and the same concentration of soot in the entire flame, which is not realistic in a larger flame. Emissivity and absorptivity increases with the flame thickness and are therefore more important for larger fires30. A general rule is that if the flame is thick (over 1 m) and bright yellow (such as hydrocarbon flames), the emissivity can be assumed to be 131. But for a relatively clean combustion resulting from the burning of propane, the emissivity cannot be assumed to be 1 (see chapter 3.1.2.1 about estimating the emissivity of a flame). The emissivity of the probe body depends on the chosen surface material and how it is processed.
In order to achieve the requirement that about 90% of the heat transfer will be by radiation, it is important that the flame is thick and surrounds the object completely. It is also important to enclose the object completely in flames to get the heat impact equal from all directions.
One way to calculate the incoming radiation to an object through tests is to approximate the steel temperature as the adiabatic surface temperature, TAST32
. This is possible due to the insulation of the probe. More about the measurements on sight is presented in chapter 2.2. The adiabatic surface temperature is the equilibrium temperature of a surface which does not absorb or lose any energy, i.e. ̇ . If the radiation is dominant, the convective heat flux appears cooling. Therefore, the adiabatic surface temperature is a weighted average, and will be a value between the radiation temperature (Tr) and the gas temperature (Tg). Two other important factors are the surface emissivity (ε, see chapter 3.2.1.2), which controls the amount of energy emitted from the object and the convective heat transfer coefficient (hc, see chapter 3.2.1.1) that affect the convection heat transfer to the object. With all the values known, the incident heat flux can be calculated33:
̇ (3.13)
3.1.2.1 Estimating the theoretical emissivity of the flame, εfl
A visible flame always has some form of emissivity34. The yellow color of the flame is burning carbon particles emitting radiation. Individual soot particles are assumed to be black bodies and having a surface emissivity close to 1. Therefore the concentration of soot in a flame is a measure of the extent to which energy will be lost from the flame by radiation. Flame emissivity is therefore among other proportional to the concentration of soot. Flame emissivity can be calculated with Equation 3.12. The emission coefficient (K) is affected by the soot concentration. A higher concentration gives a higher value of K, which in turn gives a higher emissivity from the flame. This leads to that a flame with a high soot concentration will have a lower temperature than the corresponding flame with low soot concentration.
29 Drysdale, An introduction to fire dynamics, 77.
30 Wickström, Heat transfer in fire technology, 93.
31 Drysdale, An introduction to fire dynamics, 79.
32 Wickström, Heat transfer in fire technology, 129.
33 Ibid, 53.
34 Drysdale, An introduction to fire dynamics, 76-77.
K = effective emissivity coefficient (3.12) L = path length of the flame
12 If different burning substances are compared, the various soot concentrations will cause very different radiation from the flames. The exponential function makes the emissivity closer to 1 when the flame becomes thicker, see Figure 3. This non-linear function makes a comparison between a flame with a small base area and a flame with a large base area difficult. As the figure below shows, a substance that gives high soot concentrations approach ε = 1 already at about 0.5 m flame thickness while a subject with lower concentration may need a thickness of over 2 meters. This must therefore be taken to account when doing small-scale comparisons of radiation from flames of different fuels.
Figure 3. Emissivity curves for different effecive emission factors35.
The effective emission coefficient must be measured by experiments and reliable values are hard to find.
3.1.2.2 Emissivity of propane and Jet A1 flames
In order to calculate the emissivity with Equation 3.12 the effective emission coefficient must be determined by experiments. For many substances there are no values documented on the emissivity constant. For propane, it is not a too rough approximation that the gas burns with a clean flame36. Ideal combustion of propane is given by:
C3H8+5 O2→ 3 CO2+4 H2O (3.14)
For ideal combustion no soot particles will occur. To calculate an approximately flame emissivity for propane ideal combustion is assumed. Then radiation from the flame then consists mostly of the radiation from the hot water and carbon dioxide gases. The emissivity for steam and carbon dioxide in different temperatures are given in Figure 4.
35 Drysdale, An introduction to fire dynamics, 69.
36 Ibid, 26.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0 0.5 1 1.5 2 2.5 3
Emissivity
Flame thickness [m]
Emissivities of flames
Polystyrene, κ = 5.3 Polypropylene, κ = 1.8 PMMA, κ = 1.3
13
Figure 4. Emissivity for water vapor and carbon dioxide37.
The total gas emissivity is obtained by:
(3.15)
From Figure 4 values for L = 0.5 m and T = 1073K (800 °C) is taken:
This gives:
(3.16)
For 1.0 m and T = 1073K (800 °C) Figure 4 gives:
This gives:
(3.17)
The calculations show that the emissivity will increase with the larger diameter. This value is likely to underestimate the emissivity of the flame. The combustion of the propane will not be ideal and soot particles will occur which will increase the radiation from the propane flames.
37 DiNenno, SFPE Handbook, 1-82.
14 For jet fuels the assumption of ideal combustion is unrealistic. Instead the effective emissivity coefficient must be determined. In experiments it has been found that the kerosene (which is the main content of Jet A1) has an emissivity of about 0.45 at a flame thickness of 0.2 m and 0.28 at a thickness of 0.1 m 38. This gives:
⇒ ⇒
(3.18) ⇒ ⇒ (3.19) To calculate the effective emission coefficient a mean is taken of the values for 0.1 and 0.2 m:
(3.20)
This gives the emissivity diagram below (Figure 5).
Figure 5. Emissivity curve of for a kerosene flame.
This shows that the emissivity of the small scale test (L = 0.5 m) would give an emissivity of about 0.8 and in full scale (L = 1 m) of about 0.95. These values are summarized in Table 1.
Table 1. Emissivity for propane and kerosene for different flame thickness.
Substance Emissivity at L = 0.5 m Emissivity at L = 1.0 m
Propane 0.38 0.5
Kerosene 0.8 0.95
38 Sudheer, Measurement of Flame Emissivity of Hydrocarbon Pool Fires, 197 0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0 0.5 1 1.5 2 2.5 3
Emissivity
Flame thickness [m]
Emissivity for kerosene (when K=3.14)
κ = 3.14
15
3.2 Measuring T
r, T
gand T
ASTThe common way to measure the temperatures in a fire is with a thermocouple (TC)39. The TC consists of two interconnected metals. When two different metals are connected they form an electrical potential that is almost linear with temperature. A major advantage of thermocouples is that they are relatively cheap and easy to replace.
The most common type of TC is the K-type element consisting of nickel with 10% chromium interconnected with nickel, 2% aluminum, 2% manganese and 1% silicon, see Appendix E. The melting point is 1400 ˚C and the sensor measures well even in high temperatures. This type of TC is also recommended in STANAG 4240.
Measuring temperatures may seem simple, but several factors will influence the results and the same tests can show wide variation in temperature depending on the method chosen. The temperature recorded by the TC is a weighted average between the gas temperature and the radiation temperature. Heat transfer coefficients for radiation and convection are used as weights.
The thermocouple temperature is given by:
(3.21)
Therefore it is important to understand what affects mainly the convective heat transfer coefficient.
The radiative heat transfer coefficient (hr) is independent of area and plume velocity and is therefore not as interesting to discuss. An important property affecting the convective heat transfer coefficient is the size of the thermocouple40. Under the same conditions, a smaller TC will have a larger hc than a larger TC. Therefore the smaller TC will measure more close to the gas temperature. If a larger TC is used instead, the convective heat transfer coefficient will decrease and the temperature will come closer to the radiation temperature. The TC will never be able to measure the exact Tr and Tg, only temperatures close to them. Simplified a small thermocouple, about 0.25 – 0.50 mm diameter, will measure the gas temperature while a larger thermocouple will be affected too much by the radiation.
A special case of TC is when providing the thermocouple with a large surface area which is insulated at the back, a so-called plate thermometer (PT)41. This is constructed by a thin steel plate with a welded thermocouple on the mid-back. The insulation at the back of the plate reduces the heat transferred from the surface to a minimum. With a PT the convective heat transfer coefficient is relatively low and the plate thermometer will measure the temperature near to the radiation temperature42. The temperature of a surface which cannot absorb any heat is called the adiabatic surface temperature. By approximating the measured steel temperature as the adiabatic surface temperature and by measure the gas temperature with a small TC the total incident heat flux to the object can be determined by43:
̇ (3.22)
39 Wickström, Heat transfer in fire technology, 116.
40 Ibid, 118.
41 Ibid, 130.
42 Ibid, 119.
43 Ibid, 53.
16 The steel surface temperature (approximated as adiabatic surface temperature) and the gas temperature are measured in the experiments but the surface emissivity and convective heat transfer coefficient will have to be estimated by calculations, see 3.2.1.
Figure 6. The steel temperature of an object affected by a constant fire temperature.
The time constant is then the time when the temperature has reached 63% of its final value.
A description of the response time of a measuring instrument is the time constant, τ44. The time constant is defined as the time needed to rise the temperature to 63% of its final value, see Figure 6.
By calculating the time constant an approximate time to equilibrium can be estimated. The time constant for a lumped heat case can be calculated by45:
(3.23)
where d is the thickness of the object, c is the specific heat capacity and ρ is the density of the material. To estimate the time constant the radiation temperature and convective heat transfer coefficient must be known. Because of the variation of hr this calculated time constant will be an underestimate of the real time constant.
3.2.1 Estimating the heat transfer coefficient and the surface emissivity
To calculate the incident heat flux to an object the surface emissivity and the convective heat transfer coefficient must be estimated. The surface emissivity is relatively simple, but the convective heat transfer coefficient is more complex. The estimations are not dependent of the burning substance, but rather to the heat release rate and the flow velocity in the flame. Therefore the same estimations can be used in both the jet fuel fire and the propane fire (if the heat release rate and the flow rates are the same).
3.2.1.1 The heat transfer coefficient, hc
The convection around a thermocouple or a probe in a flame is not created because of the temperature difference between the gas temperature and the object’s surface temperature.
Therefore the convection can be assumed the same as a forced convection (like the convection when
44 Wickström, Heat transfer in fire technology, 119.
45 Ibid, 166.
17 a fan is blowing on an object)46. In all convection problems a boundary between the flowing fluid and the surface arises. It is this boundary layer and its properties that have a major impact on the convective heat transfer. Simplified the thickness of the boundary layer and the thermal conductivity of the fluid controls how much heat is transferred from the gas to the surface47. The thickness itself then depends on the velocity of the fluid.
To calculate the heat transfer coefficient is therefore a complex equation depending on several factors. In order to describe the fluids behavior around the object the Reynolds number can be calculated. A Reynolds number of 2 x 105 indicates laminar flow and above 3 x 106 the flow generally is turbulent48. Reynolds number for a round body is given by49:
(3.24)
Table 2. Constants used in Equation 3.25 based on Reynolds number, from Wickström, Heat transfer in fire technology, 64.
Calculated from Equation 3.24, Table 2 gives the constants A, C and n which are used in Equation 3.25 below:
(3.25)
The flow velocity in the flame, , can be calculated with McCaffrey:s plume model50. This model is developed by experiments and can be used both in a continuous flame, during the transition from flame to plume and in the plume over the fire. For calculations in the flame:
( ̇ ) ̇ where (3.26)
46 Ibid, 60.
47 Drysdale, An introduction to fire dynamics, 56.
48 Ibid, 55.
49 Wickström, Heat transfer in fire technology, 64.
50 Karlsson, Enclosure fire dynamics, 79.
u = velocity in the flame centerline z = height where the velocity is measured ̇= heat release rate from the fire in kW
18 To calculate the centerline velocity in the plume the heat release rate of the fire ( ̇) is needed. The heat release rate can be calculated, but certain material properties are needed. More about this is presented in 3.3. The heat release rate can be calculated by:
̇ ̇ where (3.27)
3.2.1.2 The surface emissivity, ε
The surface emissivity is tabulated in a variety of sources. For example the emissivity of polished steel can be found to differ from 0.14 to 0.38 51. To ensure that the probe has the same emissivity during the whole test (and during different tests), the probe is burnt in about 800 – 1000 ˚C for several hours. The probe is then covered with a black surface with the emissivity ε = 0.9 which will not change during the test.
3.3 Difference between Jet A1 and propane flame
A flame can only occur in a gas. While a liquid fuel has to evaporate first52, the gas can be burned at once. For combustion an oxidizer is needed, which usually is oxygen from the air. There are two ways for a fuel in the gas phase to burn; one way is that the gas is premixed with oxygen (or air) before combustion. The second way is when the gas from the beginning is separated from the oxygen, but will burn in the area where it has reached the correct oxygen/gas mixture53. These give rise to premixed and diffusion flames respectively. The latter is encountered in the combustion of gas jets, flammable liquids and solids54.
Figure 7. Different ways of heat transfer in a pool fire, from Drysdale, An introduction to fire dynamics, 13.
51 Drysdale, An introduction to fire dynamics, 62.
52 Ibid, 2
53 Ibid, 12
54 Ibid, 12
A = area of the fire ̇ = mass loss rate =combustion efficiency = heat of combustion
19 The burn rate of a diffusion flame can be calculated using the mass flow of the combustible gas, which for a gas jet flame is independent of the combustion process. For flammable liquids and solids, the mass loss rate is directly dependent on the heat transfer from the flame to the fuel. The rate of burning ̇ can generally be expressed as:
̇ ̇ ̇ (3.28)
where ̇ is the heat flux from the flame (W/m2) and ̇ represents the losses expressed as a heat flux through the fuel surface (W/m2). LV is the heat required to evaporate the fuel (J/kg), which for liquids is its latent heat of evaporation. The heat flux ̇ depends on the energy released within the flame and the heat transfer involved55.
The most important factor that characterizes a flames behavior is the rate at which energy is released ( ̇), it is given by the expression:
̇ ̇ (3.29)
where is the area of the fuel surface (m2), is the heat of combustion of the volatile (J/kg), ̇ is the mass loss rate (kg/m2s) and χ is a factor (<1.0) to compensate for incomplete combustion.
This can be summarized that a gas (such as propane) do not need to vaporize before combustion, while a liquid fuel (such as Jet A1) does. This means that there will be heat losses in the combustion of a liquid fuel due to the energy required to evaporate the liquid to gas.
55 Drysdale, An introduction to fire dynamics, 13
20
4 Results
4.1 Sand bed burner
The sand bed burner is constructed in modules to make the construction more mobile. STANAG 4240 specifies that the base of any burner must be at least 1 meter wider on each side of the object. Even the UN document “Classification procedures”56 provides 1 meter on each side of the object as a minimum burner area.
To get a thicker flame than with a premixed flame the gas is diffusing through a sand bed to form a diffusion flame (see chapter 3.3). The gas is first spread over the entire burner area in an air gap under the sand bed. The resistance is greater in the upper layer which will fill the whole lower layer with gas before it diffuses through the sand. In this way the gas flow area is as big as possible through the upper sand bed.
4.1.1 Construction of the sand bed burner
The mission statement from Bofors Test Center (see Appendix C) says that the burner is to be constructed for a maximum test object size of 0.16 x 0.87 m. Therefore the total minimum size of the sand bed burner must be 2.16 x 2.87 m (1 m on each side). In order to use the current test table of Bofors in small scale one module must be less than 1.24 x 1.13 m. The proposed plan means that the burner is divided into six interlocking modules of 1.15 x 1.0 m each. This gives a total area of 2.3 x 3.0 m.
To separate the air gap and the sand layer a perforated steel sheet on a steel frame is used. The frame is made to just fit the inner dimensions of the burner. The sand is selected to be of particle size 4-8 mm. The holes in the perforated steel plate must therefore be max 3 mm diameter. The total hole area of the sheet is then 51% of the sheet area. The height of the sand layer is chosen to 100 mm. The flow of propane required to achieve the same heat release rate as a Jet A1-test is about 0.044 kg/m2s, see Appendix D.
This construction means that only one inlet for gas is needed. The weight of the structure is assumed to be sufficient for the burner modules to remain on their places during the tests. Therefore no hardware for interconnection between the modules will be needed.
Figure 8. Principle design of the sand bed burner. All measurements are in millimeters. For design drawings, see Appendix A.
56 United Nations, Part I Classification procedures.
21
Figure 9. Pictures of the sand bed burner used for the tests.
In this project only one module will be made. The possibility to build a large burner for real tests will be evaluated. Design drawings of the sand bed burner are presented in Appendix A.
4.2 Adiabatic surface temperature probe
There are many factors that come into play when the impact of different burning substances shall be compared. The main factors for the heat transfer is the temperature of the flame, the radiation from the flame, convective heat, convective heat transfer coefficient and specimen management properties and surface characteristics. For these features to be exactly equal it is necessary to burn the same substance and use the same specimen. The only heat transfer of theoretical interest in insensitive munition tests is the heating of the explosive charge in the munition, i.e. the total energy input into the specimen. If this is done by radiation or convection should not be a significant factor.
It’s therefore of interest to compare the total energy supplied to the specimen. One way is by building a probe that measures the steel surface temperature and the gas temperature. The steel surface temperature is then approximated as the adiabatic surface temperature. With these values, the heat transfer is calculated; see chapter 3.2. By using the same probe in the tests of kerosene and LPG, steel surface temperature and gas temperatures are recorded. This makes the different heating processes comparable.
To get an as accurate measuring instrument (probe) as possible it shall have these four properties57: 1. a similar form as the actual specimen
2. similar surface properties as the actual specimen 3. an insulated surface
4. have a short response time
The first property refers to convective heat transfer, the second one refers particular to emissivity which is crucial for the heat transfer by radiation. Ideally, the surface should be perfectly insulated, which is impossible. The fourth property is important for transient problems.
4.2.1 Construction of the probe
Since the test objects mostly are cylindrical the probe will be designed in a similar manner. The probe is made of a steel tube with a wall thickness of 2 mm and insulated with fire insulation. The probe measures the steel surface temperature and the gas temperature. The steel temperature is then approximated as the adiabatic surface temperature. In this way, the incoming heat flux to a
57 Wickström, Heat transfer in fire technology, 135
22 specimen can be determined. At the probe surface there will be four welded ø 0.8 mm TC measuring the steel surface temperature. The best way to weld the thermocouples is on the inside of the tube to protect the wires from direct heat and for an easier installation. But if this is not possible, holes can be drilled where the TCs can be passed through, and then welded to the outside. When the sheet is thin, it can be assumed that the temperature is uniform58. Just outside the probe surface (about 20 mm) there are four ø 0.25 mm TC which purpose is to measure the gas temperature. For product specifications, see Appendix E. To obtain an emissivity of about 0.9 at the pipe surface the pipe is burned over a long time (12 – 24 hours) at about 800 – 1000 °C before use. The emissivity will then be constant throughout the experiment.
Figure 10. Principle drawing of the probe.
Design drawings of the TAST-probe are presented in Appendix B.
Figure 11. Pictures of the probe used in the tests.
58 Wickström, Heat transfer in fire technology, 21
23 4.2.2 Test of the probe
The test of the probe was made in the fire testing oven in the laboratory at Luleå University of Technology. The oven was heated according to the ISO 834-curve up to 830 ˚C and then maintained at this temperature for about 5 minutes. The heating process was then stopped and the oven was cooled off. The test was repeated with the probe rotated 180˚ to verify that the heating pattern is similar with different thermocouples in different positions. The total test time was about 70 minutes.
The figures below show the steel surface temperature and the mean gas temperature during test 1 and test 2. The measured values are compared with the ISO 834-curve and the plate thermometer in the fire oven.
The oven is programmed to follow the ISO-curve, but the heating process is divided in 4 time steps.
Each step has a linear approximation to the ISO-curve. Therefore, the heating will not be completely according to ISO-curve, but instead be divided in 4 linear steps. This can be seen by looking at the figures including the ISO-curve and the gas and PT temperatures, see for example Figure 14.
Figure 12 shows the steel surface temperature during the first test. The TC placed nearest the flame is the bottom TC, TC 2 (down). This TC is also the measuring point showing the highest temperature.
The measuring point furthest away from the flame (also shaded from radiation from the flame), TC 4 (up), is the one that shows the lowest temperatures. This pattern is also repeated when the probe is rotated during test 2, see Figure 13.
During test 1 the temperature fell due to a technical problem when the ISO-curve was interrupted and the temperature kept constant at about 800 ˚C. Therefore all curves decline for a few seconds after approx. 1600 seconds, see Figure 12.
Figure 12. The steel surface temperature during the first test, compared with the plate thermometer in the oven and the theoretical ISO-curve.
0 100 200 300 400 500 600 700 800 900
0 500 1000 1500 2000 2500 3000
Temperature [˚C]
Time [s]
Probe surface temperature during test 1
ISO-curve PT in oven
TC 1 (right) TC 2 (down) TC 3 (left) TC 4 (up)
24
Figure 13. The steel surface temperature during the second test, compared with the plate thermometer in the oven and the theoretical ISO-curve.
Figure 14 shows the mean gas temperature during test 1. During the most intensive heating the gas temperature lies clearly over the PT in the oven. During the rest of the test the mean gas temperature is approx. similar to the PT temperature. The same pattern is shown during test 2, see Figure 15.
Figure 14. The mean gas temperature in the oven during the first test, compared with the plate thermometer in the oven and the theoretical ISO-curve.
0 100 200 300 400 500 600 700 800 900
0 500 1000 1500 2000 2500 3000 3500
Temperature [˚C]
Time [s]
Probe surface temperature during test 2
ISO-curve PT in oven
TC 1 (left) TC 2 (up) TC 3 (right) TC 4 (down)
0 100 200 300 400 500 600 700 800 900
0 500 1000 1500 2000 2500 3000
Temperature [˚C]
Time [s]
Mean gas temperature during test 1
Mean gas temp PT in oven ISO-curve
25
Figure 15. The mean gas temperature in the oven during the second test, compared with the plate thermometer in the oven and the theoretical ISO-curve.
4.3 Test results
The results from the three tests conducted at Bofors Test Center are presented below.
4.3.1 Sand bed burner test
The test with the sand bed burner was performed early in the morning to minimize the influence of wind. The first test was performed to measure the temperatures at different heights. The results are presented in Figure 16.
Figure 16. Temperatures measured at different heights with a coarse 3 mm TC.
0 100 200 300 400 500 600 700 800 900
0 500 1000 1500 2000 2500 3000
Temperature [˚C]
Time [s]
Mean gas temperature during test 2
Mean gas temp PT in oven ISO-curve
0 100 200 300 400 500 600 700 800 900
0 50 100 150 200 250
Temperature [˚C]
Time [s]
Temperatures at different heights
20 cm 40 cm 60 cm 80 cm
26 The second part is the actual test with the probe. The measurement was chosen to be made 200 mm over the sand bed, see Figure 17. The test lasted for 395 seconds (approx. 6.5 minutes). Mean gas flow during the test was 0.038 kg/s (compared to the 0.044 kg/s needed, see Appendix D). The results from the test are presented in Figure 18 and Figure 19 below. Both graphs show the temperatures in all four directions.
Figure 17. Test setup for the sand bed burner test.
27
Figure 18. The surface temperature of the probe for all four directions during the test with the sand bed burner.
Figure 19. The gas temperature in all four directions during the test with the sand bed burner. Measured with a 0.25 mm diameter TC. The graph is not for reading certain values, but to see trends.
During the test one of the 10 tubes was weighed to be able to analyze variations in the gas flow. The weight of the tube was noted every 5th second, see Figure 20. The flow pattern from tube 10 is assumed to be representative for the gas flow from all tubes.
0 100 200 300 400 500 600 700 800
0 50 100 150 200 250 300 350 400
Temperature [˚C]
Time [s]
Probe surface temperature
Down Left Up Right
0 200 400 600 800 1000 1200
0 50 100 150 200 250 300 350 400
Temperature [˚C]
Time [s]
Gas temperature
Down Left
Up Right
28
Figure 20. The gas flow from bottle 10 during the test with the sand bed burner.
Every tube was weighed before and after the test to calculate a mean value of the gas flow. The measurements are presented in Table 3 below.
Table 3. Weight of tubes before and after the test.
Tube Before [kg] After [kg] Difference [kg]
1 14.86 13.12 1.74
2 14.38 13.17 1.21
3 13.9 12.64 1.26
4 12.89 11.82 1.07
5 13.08 12.34 0.74
6 14.44 12.68 1.76
7 12.56 11.22 1.34
8 14.28 12.42 1.86
9 14.53 12.24 2.29
10 14.11 12.49 1.62
Sum: 139.03 124.14 14.89
The test lasted for 394 seconds. The mean gas flow during the test is calculated to:
(4.1)
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009
0 50 100 150 200 250 300 350 400
Gas flow [kg/s]
Time [s]
Gas flow from tube 10 during SBB test
Gas flow
29 4.3.2 Small scale Jet A1-test
The test with Jet A1 was made after the test with the sand bed burner. Wind speed was still low, but some gusts of wind occurred. The measurement was made 200 mm over the edge of the vat. The test lasted for 600 seconds. 30 liters of Jet A1 and 4.5 liters of flight petrol were consumed during the test. The results from the test are presented in Figure 21 and Figure 22 below. Both graphs show the temperatures in all four directions.
Figure 21. The surface temperature of the probe for all four directions during the test with the Jet A1 pool fire.
Figure 22. The gas temperature in all four directions during the test with the Jet A1 pool fire. Measured with a 0.25 mm diameter TC. The graph is not for reading certain values, but to see trends.
0 100 200 300 400 500 600 700 800 900
0 100 200 300 400 500 600
Temperature [˚C]
Time [s]
Probe surface temperature
Down Left Up Right
0 200 400 600 800 1000 1200
0 100 200 300 400 500 600
Temperature [˚C]
Time [s]
Gas temperature
Down Left
Up Right
